The results of whole-genome CRISPR–Cas9 screening in SLC7A11-overexpressing 786-O cells under glucose-replete and -starved conditions were downloaded from previous study (2).

The log2 (TPM + 0.001) transformed normalized gene and transcripts expression profiles, copy number variations on gene expression estimated using the Genomic Identification of Significant Targets in Cancer 2.0 (GISTIC2.0) method, DNA methylation profiles and phenotype data of pan-cancer were download from the University of California, Santa Cruz Genome Browser (UCSC) Xena database (https://xenabrowser.net/). Moreover, the proteomics data of multiple cancer types were obtained from the Clinical Proteomic Tumor Analysis Consortium (CPTAC) database (https://proteomics.cancer.gov/programs/cptac). The ProteoCancer Analysis Suite (PCAS) package was used to analyze the CPTAC datasets (11). Also, several datasets for various cancer types were obtained from GEO datasets (https://www.ncbi.nlm.nih.gov/geo/).

To understand the biological and clinical significance of RPN1, correlations were analyzed between the expression of RPN1 and genesets expression, Tumor Mutation Burden (TMB), immune cell infiltration, and Tumor Immune Dysfunction and Exclusion (TIDE) score, as well as drug sensitivity score. Oncogenes were obtained from ONGene (http://www.ongene.bioinfo-minzhao.org) (12) databases. A total of 11 immune checkpoint genes (ICGs) were extracted from previous studies (13). Immune cell infiltration scores came from the TIMER2.0 database (http://timer.cistrome.org/) (14). TIDE scores were predicted using the online TIDE tool (http://tide.dfci.harvard.edu/) (15). Data from the Genomics of Drug Sensitivity in Cancer database (GDSC, https://www.cancerrxgene.org/) was used to assess drug sensitivity using “oncoPredict” (16). Spearman’s method (“psych” package) was used for correlations.

Gene set enrichment analysis (GSEA) was performed using the “ClusterProfiler” R package to explore gene set enrichment based on correlation analysis. Specifically, two gene set collections were employed: c2.cp.kegg.v2023.2.entrez.gmt for signaling pathway analysis and c5.go.bp.v2023.2.entrez.gmt for biological processes associated with RPN1. This analysis spanned multiple cancer types to identify consistently enriched pathways and biological processes. Following GSEA for each cancer type, commonly enriched gene sets were extracted and visualized to underscore their significance.

RPN1 shRNA sequences were designed using the BLOCK-iT™ RNAi Designer tool (Thermo Fisher Scientific, URL: https://rnaidesigner.thermofisher.com/rnaiexpress). The annealed shRNA double-stranded fragments were cloned into the pGreen vector. After assessing the knockdown efficiency of multiple candidate shRNAs, two shRNAs targeting RPN1 were selected for subsequent experiments. Additionally, a scrambled non-specific control shRNA (shNC) was cloned into the same vector and used as a negative control.

Human cancer cell lines, including A549 and H1299 (lung cancer), T24 (bladder cancer), SW480 (colon cancer), MDA-MB-231 (breast cancer), MGC-803 (gastric cancer), and U87 (glioblastoma), were acquired from the American Type Culture Collection (ATCC). Cells were cultured in DMEM supplemented with 10% FBS at 37°C in a 5% CO₂ atmosphere. After a 24-hour incubation period, lung cancer cells were transfected with 2.5 μg of shRNA using Lipofectamine 6000 reagent (Beyotime, China) according to the manufacturer’s instructions.

For the EdU assay, cells were treated with 10 μM EdU for 2 hours, fixed with 4% paraformaldehyde, and permeabilized with 0.3% Triton X-100 in PBS. The Click reaction solution from Beyotime Institute of Biotechnology (China) was used for subsequent processing. Following a 24-hour incubation, cell images were captured using an inverted fluorescent microscope and analyzed using NIH ImageJ software (Version 1.8.0).

For the CCK-8 assay, cells were incubated with the CCK-8 working solution for 1-4 hours at 37°C. Absorbance was measured at 450 nm using a microplate reader, and the absorbance values were used to calculate cell viability percentages relative to control samples.

Total RNA was extracted from cultured cells using a commercially available RNA isolation kit, following the manufacturer’s protocol. RNA concentration and purity were assessed using a spectrophotometer (NanoDrop 2000, Thermo Fisher Scientific, USA). cDNA synthesis was performed using a reverse transcription kit (cDNA Reverse Transcription Kit, Thermo Fisher Scientific, USA) according to the manufacturer’s instructions. For qPCR, we used specific primers for CDKN1A (Forward: TGTCCGTCAGAACCCATGC, Reverse: AAAGTCGAAGTTCCATCGCTC) and CDKN2A (Forward: GGGTTTTCGTGGTTCACATCC, Reverse: CTAGACGCTGGCTCCTCAGTA), with a reference gene, ACTB (Forward: CATGTACGTTGCTATCCAGGC, Reverse: CTCCTTAATGTCACGCACGAT). The PCR reactions were performed on a real-time PCR system (ABI Q6, Applied Biosystems, USA), using SYBR Green assay. The thermal cycling conditions were as follows: 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec, 60°C for 1 min. Relative expression levels were calculated using the comparative Ct method (ΔΔCt).

Senescent cells were identified using a fluorescein-based probe targeting β-galactosidase activity. The probe, containing two galactoside moieties, is cleaved by β-galactosidase within lysosomes under acidic conditions, emitting a fluorescent signal with absorption/emission maxima of 490/514 nm. Post-incubation with the probe, cells were fixed and nuclei counterstained with DAPI. Fluorescent signals were visualized using a fluorescence microscope equipped with standard filter sets.

Protein expression levels of Ki67, P21, and P16 were assessed by Western blotting, with ACTB serving as the internal control. Protein samples were resolved by SDS-PAGE, transferred to PVDF membranes (Millipore, USA), and incubated with primary antibodies: Ki67 (Catalog No. 27309-1-AP, 1:1000 dilution), P21 (Catalog No. 10355-1-AP, 1:1000 dilution), P16 (Catalog No. 10883-1-AP, 1:1000 dilution), sourced from Proteintech (Rosemont, IL, USA). Membranes were then incubated with HRP-conjugated secondary antibodies (Catalog No. SA00001-2/1, 1:5000 dilution, Rosemont, IL, USA) at a 1:5000 dilution for 1 hour at room temperature. Bands were visualized using an enhanced chemiluminescence kit (Beyotime, China). The band intensity was quantified using ImageJ software (NIH, Bethesda, MD, USA), normalized against ACTB levels for accurate comparison across samples.

ELISA was conducted to measure IL6 and IL8 concentrations in the culture supernatants of RPN1 knockdown cells. ELISA kits for IL6 and IL8 were purchased from R&D Systems (Minneapolis, MN, USA). Results were presented as mean values for shRPN1 and shNC groups, with standard curves generated for each cytokine to ensure accurate quantification. Data analysis was performed using GraphPad Prism software (GraphPad Software, San Diego, CA, USA) to compare cytokine levels between the RPN1 knockdown and control conditions.

All animal experiments adhered to institutional animal ethics guidelines and were conducted following the ethical standards approved by the Laboratory Animal Ethics Committee of the Experiment Animal Center of Soochow University (approval no. 202401A052). Six nude mice per group were used: one group received subcutaneous injections of H1299 cells transfected with shNC, while the other group received shRPN1-transfected H1299 cells. Tumor growth was monitored weekly by recording body weight and tumor volume.

At the endpoint of the study, all animals were euthanized humanely under deep anesthesia to minimize suffering. Anesthesia was achieved using an intraperitoneal injection of pentobarbital sodium at a dose of 50 mg/kg. Following confirmation of deep anesthesia, as indicated by the loss of the righting reflex and absence of response to a toe pinch, euthanasia was performed by cervical dislocation. After euthanasia, tumors were promptly excised, weighed, and analyzed. Tumor volumes were calculated using the formula V = (length × width²)/2. Body weight and tumor size data were used to evaluate the impact of RPN1 knockdown on tumor growth.

Tumor tissues were processed into paraffin-embedded sections and subjected to standard immunohistochemistry protocols. Sections were incubated with primary antibodies against RPN1(Catalog No. 12894-1-AP, 1:200 dilution), SLC7A11 (Catalog No. 26864-1-AP, 1:200 dilution), P16 (Catalog No. 10883-1-AP, 1:200 dilution), Ki67 (Catalog No. 27309-1-AP, 1:200 dilution), and CD31 (Catalog No. 11265-1-AP, 1:200 dilution), all from Proteintech (Rosemont, IL, USA). Quantitative analysis was conducted using ImageJ software (NIH, Bethesda, MD, USA) to determine the Area of Differential Staining (AOD) for RPN1, SLC7A11, and P16, the proliferation index for Ki67-positive cells, and the blood vessel areas for CD31-positive cells.

Copy number analysis was performed using qPCR with an absolute quantification method. DNA samples were extracted from cells and subjected to qPCR using specific primers targeting RPN1 (Forward: GGCCAAGATTTCAGTCATTGTGG, Reverse: CTTCGTTGGATAGGGAGAGTAGA). Standard curves from known copy number standards quantified absolute copy numbers per cell. Results were expressed as copy numbers per cell, with a copy number of 2 serving as the baseline control. All qPCR reactions were performed in triplicate, and data were analyzed using the relative quantification method.

Genomic DNA was extracted using a DNA Purification Kit, and 1 μg of DNA was modified with sodium bisulfite using the Epitect Bisulfite kit (Qiagen, Inc.) according to the manufacturer’s specifications. MSP was conducted in a 20 μl reaction volume using AmpliTaq Gold (Applied Biosystems; Thermo Fisher Scientific, Inc.). MSP products were separated on 2% agarose gels containing GelRed^® Nucleic Acid Gel Stain. Successful bisulfite modification was confirmed by amplifying the GAPDH promoter region with modified DNA samples. Primers used were: MF: 5’-AATTGTTATGTTGTTTATTTTTCGA-3’, MR: 5’-ACTACGACCTACCTTTATATACGAA-3’, UF: 5’-AATTGTTATGTTGTTTATTTTTTGA-3’, UR: 5’-AACTACAACCTACCTTTATATACAAA-3’.

Using TFTF R package to predict upstream transcription factors regulating RPN1 (17), a total of 7 transcription factor target gene prediction datasets were included, including KnockTF (18), hTFTarget (19), ChIP_Atlas (20), GTRD (21), and ENCODE (22), as well as prediction results calculated using FIMO algorithm based on core transcription factors from JASPAR database (23, 24). Further calculate the expression correlation between RPN1 and SP1 based on the pan organizational dataset of TCGA and GETx databases.

To investigate the regulatory effect of the transcription factor SP1 on the target gene RPN1, a dual-luciferase reporter assay was performed. The complete promoter sequence of RPN1 was amplified and cloned into a firefly luciferase reporter vector (pGL3-basic, Promega, Madison, WI, USA). The control vector containing a Renilla luciferase gene was used for normalization of transfection efficiency (pRL-TK, Promega, Madison, WI, USA).

Cells were co-transfected with the RPN1 promoter-firefly luciferase construct and the Renilla luciferase control vector, along with or without a SP1 expression vector. After 48 hours, the cells were lysed, and luciferase activities were measured using a Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA) according to the manufacturer’s instructions. The firefly luciferase activity was normalized to the Renilla luciferase activity to determine the relative promoter activity. The fold change in luciferase activity in the presence of SP1 was compared to the control to assess the regulatory impact of SP1 on the RPN1 promoter.

Spearman method is used for correlation analysis. The results from in vitro experiments are presented as means ± standard deviation, and were analyzed using SPSS 22.0 software (IBM Corp.). Differences between groups were analyzed using Student’s t-test or one-way ANOVA as appropriate. P < 0.05 was considered to indicate a statistically significant difference.

Based on results from a whole-genome CRISPR-Cas9 screening of SLC7A11-overexpressing 786-O cells under glucose-sufficient and glucose-starved conditions, RPN1 emerged as the third-ranked suppressor hit gene (Norm Z = 4.02, Figure 1A ). Correlation analysis from the TCGA database indicated a significant positive correlation between RPN1 and SLC7A11 expression in 24 types of tumors (r > 0.3, P < 0.05, Figure 1B ). The top five tumors by this correlation were UCS (P = 0.70), PAAD (P = 0.65), CHOL (P = 0.65), COAD (P = 0.64), and DLBC (P = 0.63). Further correlation analysis demonstrated a significant positive correlation between RPN1 expression and key genes of the pentose phosphate pathway in the majority of tumors (P < 0.05, Figure 1C ). Notably, SLC3A2 expression showed a significant positive correlation with RPN1 in 27 types of tumors (r > 0.3, P < 0.05, Supplementary Figure S1A ), and TKT expression showed a significant positive correlation with RPN1 in 26 types of tumors (r > 0.3, P < 0.05, Supplementary Figure S1B ). To validate the role of RPN1 in promoting disulfide-induced cell death, RPN1 was knocked down in eight types of tumor cells ( Supplementary Figure S2 ), which were then cultured in glucose-deprived medium. Knockdown of RPN1 significantly rescued the increase in cell death induced by 8 hours of glucose deprivation ( Figure 1D ) and the upregulation of the NADP+/NADPH ratio ( Figure 1E ).

Based on the pan-cancer dataset combining TCGA and GTEx databases, RPN1 was found to be significantly upregulated in the majority of tumor types (P < 0.05, Figure 2A ). Furthermore, transcriptomic data from the CPTAC cancer dataset also showed significant upregulation of RPN1 in several tumors (P < 0.05, Figure 2B ). Additionally, high-throughput transcriptomic data for various cancer types obtained from the GEO database indicated that RPN1 is significantly upregulated in multiple tumors (P < 0.05, Supplementary Figure S3 ). Moreover, paired t-test analysis revealed that RPN1 is significantly upregulated in paired tumor samples across several cancer types (P < 0.05, Supplementary Figure S4 ). According to the Ensembl database, RPN1 has eight transcripts, including two protein-coding transcripts ( Supplementary Figure S5A ), all of which are significantly upregulated in multiple tumors based on TCGA data ( Supplementary Figure S5B ). From the CPTAC database, we observed that RPN1 protein expression is significantly higher in various cancers compared to normal tissues (P < 0.05, Figure 2C ). Importantly, further COX analysis indicated that RPN1 is correlated with OS, DSS, DFI, and PFI in various cancers, serving as a risk factor (HR > 1, P < 0.05, Figure 2D ). Using the KMplotter online tool, high RPN1 expression was associated with poorer prognosis in patients with lung, colon, and breast cancers (HR > 1, P < 0.05, Supplementary Figures S6A-C ).

Based on GSEA enrichment analysis, RPN1 was found to be significantly enriched in multiple key biological processes across various tumors ( Figure 3A ), including mitochondrial gene expression, recombinational repair, leukocyte-mediated cytotoxicity, cell cycle checkpoint signaling ( Figure 3B ), and regulation of DNA repair ( Figure 3C ). Regarding KEGG signaling pathways, enrichment analysis revealed that RPN1 is significantly associated with several critical biological functions ( Figure 3D ), primarily including proteasome ( Figure 3E ), cell cycle ( Figure 3F ), spliceosome, DNA replication, nucleotide excision repair, mismatch repair, lysosome, and protein export. To further evaluate the significance of RPN1 in cancer progression, we analyzed the correlation between RPN1 expression and oncogene expression across pan-cancer samples. The results indicated that RPN1 is significantly positively correlated with oncogene expression in most tumors (P < 0.05, Supplementary Figure S6A ). Specifically, correlation analysis showed that the oncogenes HSPA5 and TFG are significantly positively correlated with RPN1 expression in nearly all cancers (P < 0.05, r > 0.2, Supplementary Figures S7B, C ).

Based on enrichment analysis, RPN1 was found to be associated with the proteasome, prompting us to further investigate its relationship with endoplasmic reticulum (ER) stress. Correlation analysis using transcriptomic data from the TCGA database revealed that RPN1 expression positively correlates with the ER stress-related genes PERK (EIF2AK3, Figure 4A ) and ATF6 ( Figure 4B ) in multiple cancer types. Furthermore, using the PCAS package, we observed that RPN1 protein expression also positively correlates with the levels of PERK and ATF6 in various tumors ( Figure 4C ). Phosphorylation data from the PCAS package analysis revealed that RPN1 levels significantly positively correlate with the phosphorylation of IRE1α (ERN1, Figure 4D ) and PERK ( Figure 4E ). Moreover, in three cancer cell lines (A549, MGC-803, and SW480), we analyzed the expression changes of PERK and ATF6 following RPN1 knockdown. The results showed that both PERK and ATF6 were significantly downregulated ( Figures 4F, G ). At the protein level, PERK expression was significantly decreased following RPN1 knockdown ( Figure 4H ).

To further investigate the regulatory effects of RPN1 on cell proliferation and senescence, we constructed RPN1 knockdown plasmids and conducted in vitro experiments. CCK-8 assays demonstrated that knockdown of RPN1 significantly reduced the proliferative capacity of lung cancer cells A549 ( Figure 5A ), gastric cancer cells MGC-803 ( Figure 5B ), and colon cancer cells SW480 ( Figure 5C ). EdU incorporation assays visually confirmed the reduced proliferation in these cell lines, as shown by representative images ( Figure 5D ) and quantification results ( Figure 5E ). Immunofluorescence staining for β-galactosidase activity indicated increased cellular senescence following RPN1 knockdown ( Figures 5F, G ). Further correlation analysis indicated that RPN1 expression is significantly positively correlated with cell cycle-related gene expression in LUAD, STAD, and COAD ( Supplementary Figures S8A-C ). qPCR analysis demonstrated that knockdown of RPN1 led to a significant downregulation of several key cell cycle regulatory genes ( Supplementary Figure S8D ). In addition, Ki67 mRNA levels were downregulated in RPN1-depleted cells ( Figure 5H ), whereas CDKN2A ( Figure 5I ) and CDKN1A ( Figure 5J ) mRNA levels were upregulated. These findings were corroborated at the protein level by Western blot analysis, which showed decreased Ki67 and increased expression of P16 (product of CDKN2A) and P21 (product of CDKN1A) ( Figures 5K-N ). Additionally, ELISA detected elevated concentrations of IL6 and IL8 in the culture supernatants of RPN1-knockdown cells, indicating enhanced secretion of senescence-associated inflammatory cytokines ( Figure 5O ).

We demonstrated the effects of RPN1 knockdown on subcutaneous tumor formation using shNC and shRPN1 stably transfected H1299 cancer cells in nude mice ( Figure 6A ). Tumor growth curves indicated that the shRPN1 group exhibited significantly reduced tumor growth compared to the shNC group ( Figure 6B ). At the end of the experiment, the tumor weights in the shRPN1 group were markedly lower ( Figure 6C ). qPCR analysis confirmed reduced RPN1 expression in tumors from the shRPN1 group ( Figure 6D ). Dissociation of tumor tissues for EdU proliferation assays revealed inhibited proliferation of tumor cells in the shRPN1 group ( Figures 6E, F ). Immunohistochemical staining indicated significantly decreased expression of RPN1 and SLC7A11, and increased expression of P16 in the shRPN1 group ( Figures 6G, H ). Furthermore, quantitative analysis showed a reduced proportion of Ki67-positive cells in the tumors from the shRPN1 group ( Figure 6I ), along with a decreased area of intratumoral vasculature ( Figure 6J ).

Analysis of pan-cancer DNA copy number data from the TCGA database revealed an increase in RPN1 copy number in the majority of tumor types ( Figure 7A ). Correlation analysis demonstrated a significant positive correlation between RPN1 copy number and RNA expression levels in most tumors (P < 0.05, r > 0.3, Figure 7B ). The four cancer types with the highest correlation were HNSC (r = 0.675, Supplementary Figure S9A ), LUSC (r = 0.664, Supplementary Figure S9B ), ESCA (r = 0.654, Supplementary Figure S9C ), and LUAD (r = 0.609, Supplementary Figure S9D ). qPCR analysis confirmed a significant increase in the average RPN1 copy number in several cancer cell lines ( Figure 7C ). On the other hand, based on methylation array data, we identified a significant association between RPN1 expression and methylation levels at specific CpG sites ( Figure 7D ). Using the methPrimer online tool, we designed primers for MSP and identified CpG islands in the promoter region of RPN1 ( Supplementary Figure S9E ). MSP experiments in three cell lines confirmed the presence of DNA methylation modifications in the RPN1 promoter region ( Figure 7E ). As shown in Figure 7F , treatment of these cell lines with the DNA methylation inhibitor 5-Aza resulted in increased RPN1 expression, indicating that DNA methylation negatively regulates RPN1 expression.

Using the TFTF package, we identified SP1 as an upstream transcription factor of RPN1 ( Figure 8A ). Correlation analysis based on pan-cancer data from the TCGA database revealed a positive correlation between the expression of RPN1 and SP1 in the majority of cancers (P < 0.05, r > 0.3, Figure 8B ). Among them, the cancer types with the highest correlations were THYM (r = 0.736), KICH (r = 0.619), and LIHC (r = 0.582). Similarly, data from the GTEx database demonstrated a positive correlation between RPN1 and SP1 expression across various tissues (P < 0.05, r > 0.3, Figure 8C ). The tissues with the top three correlation coefficients were blood (r = 0.830), kidney (r = 0.824), and stomach (r = 0.677). qPCR analysis indicated that overexpression of SP1 increased RPN1 expression ( Figure 8D ). Dual-luciferase reporter assays confirmed the transcriptional activation of RPN1 by SP1 in T24 ( Figure 8E ), A549 ( Figure 8F ), and SW480 cells ( Figure 8G ), validating the regulatory role of SP1 on RPN1 transcription.

Immune therapy response is typically associated with specific immune cell infiltration. Using the TIMER algorithm, we obtained infiltration scores for six major immune cells across pan-cancer samples. Further correlation analysis revealed that RPN1 is significantly associated with the infiltration of various immune cells in multiple cancers ( Figure 9A ). Specifically, for myeloid dendritic cells, high correlation with RPN1 expression was observed in 10 cancer types (r > 0.2, P < 0.05, Figure 9B ). For macrophages, a high correlation with RPN1 expression was noted in 13 cancer types (r > 0.2, P < 0.05, Figure 9C ). Additionally, the correlation analysis of immune checkpoint genes (ICGs) expression indicated that RPN1 is correlated with several ICGs across various cancers ( Figure 9D ). For SIGLEC7, 12 cancer types exhibited significant correlation with RPN1 expression (r > 0.2, P < 0.05, Figure 9E ). Similarly, for SIRPA, a significant correlation with RPN1 expression was found in 12 cancer types (r > 0.2, P < 0.05, Figure 9F ). Furthermore, we used the TIDE algorithm to assess the immunotherapy sensitivity of tumor samples. Correlation analysis showed that RPN1 expression was significantly positively correlated with the infiltration of tumor-associated fibroblasts ( Supplementary Figure S10A ) and myeloid-derived suppressor cells ( Supplementary Figure S10B ) in multiple cancers. Moreover, RPN1 was significantly negatively correlated with T cell dysfunction ( Supplementary Figure S10C ) and significantly positively correlated with T cell exclusion ( Supplementary Figure S10D ) across various cancers.